Avalon 1246 – ASIC Chip Temperature Abnormal

Hard power-cycle at the PDU for a full 60 seconds — breaker off, caps drain, then power on. A soft reboot sometimes carries a stale `ECHU_128` flag across the restart; a full cold-start clears cached state and transient post-firmware-update glitches. Watch the first 15 minutes of API output after power-up to confirm whether the flag returns or clears. This alone clears roughly 10% of tickets in D-Central's Canaan queue.

Verify ambient at the intake grille with an IR gun — not room-middle. Target `<= 30 C` at the front face. `PVT_T abnormal` is a thermal-margin problem; if you are operating at `30+ C` ambient, you have no margin to lose and the flag will keep returning. Move the miner, add airflow, or crack a window before pursuing other causes.

Clear the front 30 cm of the intake grille — shelves, curtains, boxes, another miner's exhaust. The 1246 is a stacked-fan unit that chokes on restricted intake. Zero-dollar, two-minute fix that resolves a surprising percentage of `ECHU_128` tickets that appeared after the workshop was rearranged.

Re-pull `estats` 30 minutes after power-up: `echo -n '{"command":"estats"}' | nc <miner-ip> 4028`. Confirm `ECHU` is zero across all chains, or that the flag has returned. Save this snapshot — it is your post-Tier-1 baseline and it is the first thing D-Central's bench will ask for if the fix escalates to us.

Inspect the intake mesh for obstruction — paper shreds, pet hair, packing material, insect debris. Anything blocking even 10% of the mesh creates a microlocal airflow dead zone and the chips behind the obstruction flag as `PVT_T` outliers. Clean with a soft brush and re-test.

Compressor-blow the fin stack and fans. Real air compressor at 80-90 PSI, miner off, blow from the exhaust side back through the intake. Canned air does not move enough volume to clear a 1246 fin stack. Spin each fan blade by hand while blowing. Expect a visible dust cloud on pass one; do a second pass after 30 seconds to confirm.

Re-seat the flagged hashboard and inspect its ribbon and power connectors. Power off, wait 5 minutes. Label slots 0/1/2 before removal. Inspect IDC contacts for oxidation or blackening; wipe with 99% IPA on a lint-free wipe if needed. Reseat firmly until the latch clicks. A connector sitting 0.5 mm proud of seated feeds garbage thermal data to the MM3 controller.

Swap the flagged hashboard between slots. Move the flagged board to a different slot, power on, run 30 minutes, pull fresh `estats`. If `ECHU_128` follows the board, the board itself is the problem (paste, chip, or on-board NTC). If it stays in the original slot, the control path, AUC3 bus, or slot-specific wiring is the problem. This 20-minute diagnostic saves hours of guessing.

Replace the intake and exhaust fans if any are slow or grinding. 1246 stock fans run 6000-7000 RPM at full duty. A fan below 5500 RPM at 100% duty still reports `working` but produces a regional airflow deficit that shows up as `PVT_T` outliers on the chips directly behind it. Replace all fans at once — the survivors are almost as tired as the one you just replaced.

Verify PSU rail under load. Multimeter on DC, probe at the PSU-to-AUC3 connector while the miner is hashing at full nameplate. Expect ~12.0 V input rail; below 11.7 V means the PSU is tired or the circuit is undersized. A sagging rail forces on-board voltage regulators to work harder, bleeding extra heat through the PMICs and raising nearby chip temps.

Full thermal paste refresh on the flagged board. Remove the fin stack — Phillips or Torx fasteners, preserve or replace Kapton-backed thermal pads on the PCH and voltage-domain ICs. Clean old paste with 99% IPA and lint-free wipes, no residue on die or heatsink. Apply Arctic MX-6 or Thermal Grizzly Kryonaut in a thin uniform layer, rice-grain blob per die, let mounting pressure spread it. Reassemble with even torque. This alone restores full thermal margin on the vast majority of 18-24 month old 1246 boards flagging `ECHU_128`.

Full thermal paste refresh on all three boards when diagnosis points to global paste degradation (all chains flagging). Don't do one board at a time — factory paste ages at roughly the same rate across the miner, and a single-board refresh means you'll be back in six months for the other two.

Replace crumbled thermal pads on PCH and voltage-domain ICs while the fin stack is off for paste refresh. Crumbled pads transfer heat poorly and raise nearby chip temps indirectly through the PCB. Match pad thickness with a caliper measurement before sourcing replacements — Canaan does not publish pad specs.

Reflow the repeat-offender chip only if Tier-3 paste refresh was clean and the same chip position is still flagging `ECHU_128` 2-4 weeks later. Preheat bottom at 150 C, top-side hot air at 310-330 C for ~30 s, natural cool-down, fresh paste on reassembly. The A3206 BGA tolerates one reflow well. A second reflow on the same chip within 90 days rarely holds — replace the chip instead.

Tune AUC3 IIC bus if Step 8 showed a slot-specific flag. Edit config to `--avalon7-aucspeed 200000` (down from 400000 default), leave `--avalon7-aucxdelay 19200`. A slower bus is more tolerant of marginal cables and NTC noise. Watch the log for `CODE_MMCRCFAILED`. If those events disappear and `ECHU_128` clears, you had an AUC3 bus-margin problem masquerading as a chip thermal fault.

Stop DIY and book a D-Central ASIC Repair slot when any of these are true: same chip position flags after a clean paste refresh and one reflow cycle; a reflowed chip re-flags within 30 days; visible capacitor bulging, cracked MLCCs, or burnt-component smell; thermal camera isolates a single chip running `15 C+` hotter than neighbours and paste refresh did not resolve it; slot-specific flag persists after AUC3 replacement. At that point it is bench territory — test fixture, chip-level tools, graded A3206 replacement stock. Book at d-central.tech/services/asic-repair/ — 5-10 business day turnaround, Canada / US / international.

D-Central bench process on a 1246 `ECHU_128` case: test-fixture boot with programmable load, chip-by-chip thermal isolation via API `PVT_T` extraction plus IR thermal mapping, NTC and PMIC cross-validation. Single chip confirmed as root cause gets chip replacement with graded A3206 stock, proper reflow profile, and 24-hour burn-in at nameplate before return shipping. PMIC or voltage-domain root cause gets component-level repair with matching-spec parts, not a blanket board swap.

Ship safely. Hashboards in anti-static bags, double-boxed with `>=5 cm` of foam on every side. Include a physical note inside the box with: full `estats` baseline from Step 4, firmware version, install date, last paste-refresh date, ambient at install, and which Tier 1-3 steps you've completed. Every minute our bench spends reconstructing the fault history adds to the repair bill. A well-documented ship-in saves real dollars.